STAT R Overview. R Intro. R Data Structures. Subsetting. Graphics. January 11, 2018

Transcription

1 January 11, 2018

2

3 Why use R? R is: a free public domain implementation of S, the standard among (academic) professional statisticians, available for Windows, Mac, and Linux, an object-oriented and functional programming structure, and designed to connect to high-performance programming languages like C and Fortran.

4 Why use R? R has: an open-software environment with a large community that makes getting help easy, a massive set of packages for statistical modeling, data science, visualization, and importing and manipulating data, powerful tools for replicating and communicating your results, an interactive development environment (R Studio) tailored for interactive data analysis and statistical programmming, available for Windows, Mac, and Linux, and an object-oriented and functional programming structure.

5 Reading Data files The ability to datasets into R is an essential skill. For this class, most of the files will be on the course webpage and can be directly downloaded using read.csv. Consider a dataset available at: stat408/datasets/seattlehousing.csv Seattle <- read.csv( ' stringsasfactors = F)

6 Viewing Data files A common function that we will use is head, which shows the first few rows of a data frame. head(seattle) ## price bedrooms bathrooms sqft_living sqft_lot floors waterfront ## ## ## ## ## ## ## sqft_above sqft_basement zipcode lat long yr_sold mn_sold ## ## ## ## ## ##

7

8 Data structure Overview R has four common types of data structures: Vectors Matrices (and Arrays) Lists Data Frames

9 Data structure Overview The base data structures in R can be organized by dimensionality and whether they are homogenous. Dimension Homogenous Heterogenous 1d Vector List 2d Matrix Data Frame no d Array

10 Vector Types There are four common types of vectors: logical, integer, double (or numeric), and character. The c() function is used for combining elements into a vector dbl <- c(1,2.5,pi) int <- c(1l,4l,10l) log <- c(true,false,f,t) char <- c('this is','a character string')

11 Vector Types They type of vector can be identified using the typeof() function. Note that only a single data type is allowed. typeof(dbl) ## [1] "double" comb <- c(char,dbl) typeof(comb) ## [1] "character" comb ## [1] "this is" "a character string" "1" ## [4] "2.5" " "

12 Exercise: Vectors Create a vector with your first, middle, and last names.

13 Solution: Vectors 1 Create a vector with your first, middle, and last names. andy.names <- c("andrew","blake","hoegh") andy.names ## [1] "Andrew" "Blake" "Hoegh"

14 Data Frame Overview A data frame: is the most common way of storing data in R is like a matrix with rows-and-column structure; however, unlike a matrix each column may have a different mode in a technical sense, a data frame is a list of equal-length vectors. df <- data.frame(x = 1:3, y = c('a','b','c')) kable(df) x y 1 a 2 b 3 c

15

16 ## [1] Vector : I allows you to extract certain elements from a data frame or vector (or matrix, array, list). num.vec <- seq(from = 1, to = 9, by = 1); num.vec ## [1] num.vec[1:3] ## [1] num.vec[c(1,5,8)] ## [1] num.vec[-5]

17 Vector : II also works with logical values or expressions. num.vec[num.vec > 5] ## [1] num.vec[num.vec!= 6] ## [1] num.vec[rep(c(true,false,true),each=3)] ## [1]

18 Data Frame : I The same ideas apply to data frames, but the indices now constitute rows and columns of the data frame. df <- data.frame(x=1:3, y=3:1, z=c('a','b','c')) df[,1] ## [1] df[-1,c(2:3)] ## y z ## 2 2 b ## 3 1 c

19 Data Frame : II There are also a couple built in functions in R for subsetting data frames. df$x ## [1] new.df <- subset(df, x >1); new.df ## x y z ## b ## c

20 Exercise: 1 Create a new data frame that only includes houses worth more than $1,000, (bonus) From this new data frame what is the average living square footage of houses. Hint columns in a data.frame can be indexed by Seattle$sqft_living

21 Exercise: - Solutions 1 Create a new data frame that only includes houses worth more than $1,000,000. expensive.houses <- subset(seattle, price > ) 2 (bonus) From this new data frame what is the average living square footage of houses. Hint columns in a data.frame can be indexed by Seattle$sqft_living mean(expensive.houses$sqft_living) ## [1]

22

23 Basic Plotting in R: Scatterplot Later in the course, we will spend considerable time on graphics. For now, let s consider some of the basic functionality in R. plot(seattle$price~seattle$sqft_living) Seattle$price 0e+00 6e Seattle$sqft_living

24 Living Sqft Basic Plotting in R: labels plot(seattle$price~seattle$sqft_living, ylab='price',xlab='living Sqft') Price 0e+00 4e+06 8e

25 Basic Plotting in R: pch plot(seattle$price~seattle$sqft_living, ylab='price',xlab='living Sqft', pch=16) Price 0e+00 4e+06 8e Living Sqft

26 Basic Plotting in R: color plot(seattle$price~seattle$sqft_living, pch=16, col=rgb(0,0,.3,.3),ylab='price',xlab='living Sqft') Price 0e+00 4e+06 8e Living Sqft

27 Basic Plotting in R: title plot(seattle$price~seattle$sqft_living, pch=16, ylab='price', xlab='living Sqft',main='Price vs. Living Sqft') Price 0e+00 4e+06 8e+06 Price vs. Living Sqft

28 Price Basic Plotting in R: histogram hist(seattle$price,xlab='price', breaks='fd') Histogram of Seattle$price Frequency e+00 2e+06 4e+06 6e+06 8e+06

29 Basic Plotting in R: histogram boxplot(seattle$price~seattle$bedrooms,ylab='price', col='red', xlab='bedrooms',main='price by Bedrooms for Seattle') Price 0e+00 4e+06 8e+06 Price by Bedrooms for Seattle

30 Exercise: Basic Plot Using only the subset of homes worth more than a million dollars, create a graphic.

31 Solution: Basic Plot Price by Bedrooms for Seattle Price 1e+06 5e bedrooms For homes worth more than $1,000,000

32 Solution: Basic Plot - with Code boxplot(expensive.houses$price ~ expensive.houses$bedrooms, ylab='price', col='red', xlab='bedrooms', main='price by Bedrooms for Seattle', sub='for homes worth more than $1,000,000')